Arjun Sengupta

Power-efficient camera software @ Qualcomm | Computer Engineering @ Carnegie Mellon

Canada

About

Embedded software professional, experienced in autonomous mobile robots and on-device AI.

Experience

  • Qualcomm (Full-time · 3 yrs 6 mos)
    • Engineer
      Feb 2026 - Present · 6 mos

    • Camera Engineer
      Feb 2023 - Jan 2026 · 3 yrs

      - Worked on software/firmware-based CPU power optimizations for multimedia use cases. - Analyzed various camera features for power/performance optimizations on the next generation premium-tier Snapdragon SoCs.

  • Graduate Teaching Assistant at Carnegie Mellon University
    Aug 2022 - Dec 2022 · 5 mos

    TA for Embedded systems software engineering (18642) by Dr. Phil Koopman

  • Interim Engineering Intern (Embedded Software) at Qualcomm
    May 2022 - Aug 2022 · 4 mos

    Ported an edge detection imaging algorithm for optimal performance (under power and memory constraints) on Qualcomm's Hexagon DSP - embedded within the Snapdragon 8 Gen 1 SoC. Worked for the Multimedia Competitive Analysis (R&D) team.

  • Siemens (Bangalore, India)
    • Research Engineer
      Jul 2020 - Jan 2021 · 7 mos

      Accomplished a patent grant following the successful Invention Disclosure on "Autonomous Robot system for anomaly management" (EP4047435A1 European Patent Office: https://patents.google.com/patent/EP4047435A1/en) Continued developments on embedded and autonomy software for mobile robots

    • Research Intern - Embedded AI and Autonomous Robotics
      Dec 2019 - Jun 2020 · 7 mos

      Performed my undergraduate thesis research at the "Research in Digitalization and Automation (RDA)" unit of Siemens Technology and Services. Development highlights: • Designed and developed both – embedded and autonomy software stacks for an Nvidia Jetson-based mobile-robot using Robot Operating System (ROS) and Isaac SDK. Interfaced Jetson Nano with Intel Realsense D435 and T265 cameras. • Achieved real-time performance of V-SLAM ROS packages on the Jetson Nano by tuning parameters from realsense-ros. • Developed a Stereo-Visual-Inertial Odometry (SVIO) application for the custom robot - adapted from Nvidia Kaya’s SVIO. • Developed an accurate obstacle avoidance application for a Jetson Nano + Realsense D435 system using ‘pyrealsense’ • Developed a multi-camera person re-identification and 3D tracking system using RM-Net, accelerated on Intel NCS2 with OpenVINO. Incorporated ‘depth’ to improve tracking accuracy with Intel RealsenseD435. Inventions: • Co-authored a patent application following the successful invention disclosure on “Auditory-Visual-analytics based autonomous robot system for multi-purpose anomaly detection and localization” • Co-authored an Invention disclosure on computer-vision based food-waste management system using smartphones. - Represented Siemens at SAMHAR-Covid hackathon and won "Excellent Implementable Idea" (Top-3 teams in India)

  • Executive member, Student Council at Vellore Institute of Technology
    Aug 2019 - Aug 2020 · 1 yr 1 mo

    Official representative of the School of Electrical Engineering (SELECT), VIT-Vellore